DETAILED ACTION
Status of Claims
This communication is in response to the amendment filed 12/22/2025.
Claims 1, 8 and 12 have been amended.
Claims 7 and 14 have been cancelled.
Claims 1-6, 8-13 and 15-16 are currently pending and have been examined.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-6, 8-13 and 15-16 are rejected under 35 U.S.C. §101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1:
Claims 1-6 are directed to a method (i.e., a process), claims 8-11 are directed to a system (i.e., a machine) and claim 12-13 and 15-16 are directed to non-transitory computer readable medium (i.e., a manufacture). Accordingly, claims 1-6, 8-13 and 15-16 are all within at least one of the four statutory categories.
Step 2A - Prong One:
An “abstract idea” judicial exception is subject matter that falls within at least one of the following groupings: a) mathematical concepts, b) certain methods of organizing human activity, and/or c) mental processes.
Representative independent claim 8 includes limitations that recite an abstract idea. Note that independent claim 8 is the system claim, while claim 1 covers a method claim and claim 12 covers the matching computer readable medium.
Specifically, independent claim 8 recites:
A system for computing physician attribution, comprising:
a first computing device, the first computing device having access to electronic healthcare records (EHR) data where one or more physicians were attributed for a historic patient encounter where the one or more physicians had a higher degree of involvement in the historic patient encounter than all other physicians involved in the historic patient encounter; and
a second computing device in operable communication with the first computing device, the second computing device including a processor configured to:
access the EHR data from the first computing device,
generate a training dataset from the EHR data including predefined decisions of physician attribution,
extrapolate a plurality of variables associated with physician attribution from the training dataset, and
generate and train a machine learning algorithm based on the plurality of variables associated with physician attribution such that the machine learning algorithm is configured to iteratively determine a set of weighted predictor parameters indicative of a degree of physician involvement in a patient encounter from the plurality of variables via leave-one-encounter-out cross validation, wherein all two-way interactions between each variable of the plurality of variables is considered, and wherein the machine learning algorithm is configured to predict a physician attribution score for a physician associated with the patient encounter representative of the physician's degree of involvement in the patient encounter from subsequent EHR data based on the set of weighted predictor parameters.
The Examiner submits that the foregoing underlined limitations constitute: (a) “certain methods of organizing human activity” because computing a physician’s attributions to patients’ treatment and care by analyzing electronic health records to determine a physician’s role in the patient's treatment and care are a part of a medical workflow and determines the physician’s work performance, which relate to managing human behavior/interactions between people. These limitations constitute (b) “a mental process” because predicting that a physician’s attributions can be improved is an observation/evaluation/analysis that can be performed in the human mind or with a pen and paper. Furthermore, these limitations constitute (c) a “mathematical concept” because predicting a score for a physician associated with a patient encounter based on a degree of physician involvement in a patient encounter, a set of weighted predictor parameters, weighting value of each of the weighted predictor parameters until a rate of error for the machine learning algorithm is within a predefined threshold and using a LASSO regression model are mathematical concepts. The foregoing underlined limitations also relate to claim 8 (similarly to claim 12).
Accordingly, the claim describes at least one abstract idea.
In relation to claims 2-3 and 10, these claims merely recite specific kinds of input data, such as: claim 2 - the plurality of variables includes a physician type, a physician status, and a length of stay associated with a record of a past encounter with a physician predetermined to be properly attributed, claim 3 - the set of predictor parameters includes daily progress notes per day, long notes per day, orders per day, a length of stay, and an attending parameter and a physician and claim 10 – the regression model includes at least one of linear regression, logistic regression, polynomial regression, stepwise regression, ridge regression, LASSO regression, or ElasticNet regression.
In relation to claims 4-6, 9, 11, 13 and 15-16, these claims merely recite determining steps such as: claim 4 - performing LASSO regression to select the set of predictor parameters, claim 5 - evaluating accuracy of the machine learning algorithm by feeding the machine learning algorithm with information from a validation set defining additional predefined decisions for physician attribution and associated variables, claim 6 - models relationships between the plurality of variables to select the set of predictor parameters and refining the machine learning algorithm using a measure of error associated with a score of the attribution output, claim 9 - generate a set of predictor-parameters for the machine learning algorithm, claim 11 - optimize a predetermined accuracy threshold for the machine learning algorithm, claim 13 - finds casual effect relationships between variables of the plurality of variables, claim 15 - apply a regularization method and shrink one more coefficients of the machine learning model to Zero to improve feature selection of a set of parameter-predictors for the machine learning algorithm and claim 16 - iteratively modify a weighting value of each of the weighted predictor parameters until a rate of error for the machine learning algorithm is within a predefined threshold.
Step 2A - Prong Two:
Regarding Prong Two of Step 2A, it must be determined whether the claim as a whole integrates the abstract idea into a practical application. As noted, it must be determined whether any additional elements in the claim beyond the abstract idea integrate the exception into a practical application in a manner that imposes a meaningful limit on the judicial exception. The courts have indicated that additional elements merely using a computer to implement an abstract idea, adding insignificant extra solution activity, or generally linking use of a judicial exception to a particular technological environment or field of use do not integrate a judicial exception into a “practical application.”
The limitations of claims 1, 8 and 12, as drafted is a process that, under its broadest reasonable interpretation, covers performance of the limitations by manipulating data to make predictive models, which are ways of managing personal behavior and performing math. Also, the fact that “machine learning” is used to train the model is just “apply it”, but for the recitation of generic computer components, then it falls within the “certain methods of organizing human activity” and “Mental Process” grouping of abstract ideas. Accordingly, the claims recite an abstract idea.
The judicial exception is not integrated into a practical application. In particular, the system, first computing device, first computing device, second computing device, training database, processor, and tangible, non-transitory, computer-readable media having instructions are recited at high levels of generality (i.e., as generic computer components performing generic computer functions of receiving data/inputs, determining and providing data) such that it amounts no more than mere instructions to apply the exception using the generic computer components.
Regarding the additional limitations “machine learning”, “machine learning algorithm”, and “a machine learning model” the Examiner submits that this additional limitation amount to merely using a computer to perform the at least one abstract idea (see MPEP § 2106.05(f)). Regarding the additional limitation “input a plurality of variables ……” the Examiner submits that this additional limitation merely adds insignificant pre-solution activity (data gathering; selecting data to be manipulated) to the at least one abstract idea (see MPEP § 2106.05(g)).
Thus, taken alone, the additional elements do not amount to significantly more than the above identified judicial exception (the abstract idea). Looking at the limitations as an ordered combination add nothing that is not already present when looking at the elements taken individually. For instance, there is no indication that the additional elements, when considered as a whole, reflect an improvements in the functioning of a computer or an improvement to another technology or technical field, apply or us the above-noted implement/use to above-noted judicial exception with a particular machine or manufacture that is integral to the claim, effect a transformation or reduction of a particular article to a different state or thing, or apply or use the judicial exception in some meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is not more than a drafting effort designed to monopolize the exception (see MPEP §2106.05). Their collective functions merely provide conventional computer implementation.
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into practical application, the additional elements amount to no more than mere instructions to apply the exception using generic computer components. Mere instructions to apply an exception using generic computer component provide an inventive concept. The claims are not patent eligible.
Step 2B:
Regarding Step 2B, in representative independent claim 8, regarding the additional limitations of the system, first computing device, first computing device, second computing device, training database, processor, and tangible, non-transitory, computer-readable media having instructions, the Examiner submits that these limitations amount to merely using a computer to perform the at least one abstract idea (see MPEP § 2106.05(f)).
Thus, representative independent claim 8 and analogous independent claims 1 and 12 do not include additional elements (considered both individually and as an ordered combination) that are sufficient to amount to significantly more than the judicial exception for the same reasons to those discussed above with respect to determining that the claim does not integrate the abstract idea into a practical application.
The dependent claims no not include additional elements (considered both individually and as an ordered combination) that are sufficient to amount to significantly more than the judicial exception for the same reason discussed above with respect to determining that the dependent claims do not integrate the at least abstract idea into a practical application.
Therefore, claims 1-6, 8-13 and 15-16 are ineligible under 35 USC §101.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-2, 5-6 and 8-11 are rejected under 35 U.S.C. 103 as being unpatentable over Granson (US 11,816,539 B1) in view of Lieber (US 2014/0278493 A1).
Claim 1:
Granson discloses a method of computing physician attribution (See health care provider quality metric values in column 3, lines 32-35. Also see SurgeonCheck in column 6, lines 21-32 for physicians attributed to performed procedures for lung cancer, Y bypass and a colonoscopy.), comprising:
accessing, by a processor, historical electronic health record (EHR) data (Besides exemplary dataset of input nodes 410 shown in Fig. 4, column 11, lines 15-48 representing the clinical history of the physician, see column 5, lines 6-36 record detailing medical staff, medical imaging and historical patient treatment data. In [column 3, lines 18-36] A plurality of input data metrics from the database is input to a corresponding plurality of input nodes operated by one or more processors.);
generating, by the processor, a training dataset defining physician attribution decisions by extrapolating a plurality of variables associated with physician attribution from the historical EHR data (See trained data in column 3, lines 19-35 and column 14, lines 43-67, column 3, lines 47-52, where pre-optimized variables and column 12, lines 18-39 pre-defined rules serve as correct physician attributed processing. Historical data listing procedures performed by a physician, records of medical devices used, specialized medical images as radiology orders (column 11, lines 6-17), surgical tools used, diagnoses and complications (column 5, lines 6-37) rating or scoring procedures performed by the physician/surgeon (column 8, lines 20-35, column 9, lines 30-43) and surgical site infection rate (column 17, lines 18-53) serve as a plurality of variables associated with correct physician attribution. Also, see history of medical records in column 5, lines 6-24, Fig. 1, SurgeonCheck database 144. Besides natural language extraction, extracting a plurality of threshold values or threshold value ranges (column 9, lines 1-29), see [column 20, lines 57-67] these other machine learning algorithms extract imputed data with the automated variable selection, outlier detection and error correction processes.); and
generating, by machine learning conducted by the processor, a trained machine learning algorithm that learns learn and models decisions for physician attribution in view of the historical EHR data and training dataset such that the processor executing the machine learning algorithm is configured to predict physician attribution from subsequent EHR data (See machine learning in [column 15, lines 1-40, column 17, lines 4-17] the output data are the results of the neural network predictive model for a metric target value for the average number of cases per year per provider. The neural network process utilizes the optimized variable selection set from the Bayesian network model as the training set input.), by:
extracting a set of predictor parameters (Besides natural language extraction, extracting a plurality of threshold values or threshold value ranges (column 9, lines 1-29), see [column 20, lines 57-67] these other machine learning algorithms extract imputed data with the automated variable selection, outlier detection and error correction processes.) most predictive from the plurality of variables via leave-one-encounter-out cross validation, considering all two-way interactions between each variable of the plurality of variables (Taught as missing values within portions of a data set in [column 14, lines 6-17] Imputation utilizes a series of replacement functions where missing or outlier interval variables are replaced by the median of non-missing and non-outlier values within the data set.),
assigning a weighting value to each predictor parameter of the extracted set of predictor parameters (Besides target metric taught in column 20, lines 57-67, see predictor in column 24, lines 21-50. Also, see [column 13, lines 51-58] Learning is a process by which the free parameters (i.e., weighting assigned to links) of a neural network are adapted through a continuing process of stimulation. The type of learning is determined by the manner in which the parameter changes take place. See column 11, lines 26-48, column 17, lines 25-54 scoring a provider is based on the provider having further experience, training and knowledge as a specialist member of the Fellowship of Trained Interventional Radiology and cytopathologist are parameters.), and
iteratively modifying the weighting value of each of the extracted predictor parameters until a rate of error for the machine learning algorithm is within a predefined threshold (See column 22, lines 9-33 where literature-based metric targets serve as predefined threshold. Also, see exemplary threshold value range in column 9, lines 14-29.);
wherein the machine learning algorithm as trained, when fed with the subsequent EHR data defining a patient encounter, assigns a physician attribution score representative of a physician’s degree of involvement in the patient encounter for each physician associated with the patient encounter based on the set of predictor parameters and weighting values and improves attribution data analysis (See surgical quality rating, Surgeon check (column 7, line 35 to column 8, line 26, values objectively represent surgeon qualifications, patient symptom and recovery evaluation criteria, medical facility evaluation criteria including staff, programs and infrastructure, and cost metrics, each of these values having been calculated through the SurgeonCheck ETL 142, also see column 11, lines 15-48. See data set fed in column 17, lines 4-17 where a physician and surgical procedures are scored. Besides predicting decision alternatives in column 25, lines 3-10, see winning model in [column 29, line 66 to column 30, line 14] predictive models utilizing machine learning algorithms for the metric average number of cases per year per provider, a metric from the metric set for evaluating the Cholecystectomy procedure.).
Although Granson discloses a method of computing physician attribution, using machine learning assigning a physician attribution score representative of a physician’s degree of involvement in the patient encounter for each physician associated with the patient encounter based on a set of predictor parameters and weighting values and improves attribution data analysis mentioned above, Granson does not explicitly teach indicating that each physician’s attribution decision had a higher degree among physicians involvement in the patient encounter. Lieber teaches tracking physician attribution based on a degree of physician involvement in a patient encounter (See P0002-P0003, P0016-P0017 tracking Relative Value Units (RVU) as degrees of involvement for physician involvement in a patient encounter mentioned in P0020-P0021, P0025-P0026.) and:
wherein each attributed physician of each physician attribution decision had a higher degree of involvement in a historic patient encounter than all other physicians involved in the historic patient encounter (See P0013 and [P0027] Certain embodiments allow the comparison of the information regarding the procedures and the associated RVU or Value Unit to help further analyze the particular physician or health care provider's performance and compensation in relation to other specialties. Certain procedures or services may have different RVUs or Value Units assigned to the same procedure or service, based on the assumption that the procedure or service is more difficult for one particular specialty and not the other. Also, see peer-to-peer comparing and metrics in P0028, and Total Value Unit per physician in Fig. 2, P0059.).
Therefore, it would have been obvious to one of ordinary skill in the art of managing physician value units before the effective filing date of the claimed invention to modify the method, system and software of Granson to include indicating that each physician’s attribution decision had a higher degree among physicians involvement in the patient encounter as taught by Lieber to further refine the value of the service or procedure performed to further facilitate appropriate reimbursement and compensation mentioned in Lieber’s P0002.
Claim 8:
Granson discloses a system for computing physician attribution (See Fig. 8, health care provider analysis system.) comprising:
a first computing device, the first computing device having access to electronic healthcare records (EHR) data where one or more physicians were attributed (With proper attributes as records such as length of stay (LOS), orders per day, a referring physician, see Fig. 4 nodal inputs 410 as the average number of cases per year per the selected provider for the basis of scoring the provider mentioned in column 11, lines 15-48, column 13, lines 16-49.); and a second computing device in operable communication with the first computing device, the second computing device including a processor configured to: access the EHR data from the first computing device (See Fig. 8, health care provider analysis system including mobile device with processor 800 (Fig. 8) coupled to surgeon records (124), resources (122) and maintained records of historical patient treatment, diagnoses, complications or procedures performed (Fig. 1, Fig. 2, Fig. 3 and column 5, lines 6-24).) configured to:
generate a training dataset from the EHR data including predefined decisions of physician attribution (See machine learning in [column 15, lines 1-40, column 17, lines 4-17] the output data are the results of the neural network predictive model for a metric target value for the average number of cases per year per provider. The neural network process utilizes the optimized variable selection set from the Bayesian network model as the training set input. Besides exemplary dataset of input nodes 410 shown in in Fig. 4, column 11, lines 15-48 representing the clinical history of the physician, see column 5, lines 6-36 record detailing medical staff, medical imaging and historical patient treatment data. In [column 3, lines 18-36] A plurality of input data metrics from the database is input to a corresponding plurality of input nodes operated by one or more processors.),
extrapolate a plurality of variables associated with physician attribution from the training dataset (Besides natural language extraction, extracting a plurality of threshold values or threshold value ranges (column 9, lines 1-29), see [column 20, lines 57-67] these other machine learning algorithms extract imputed data with the automated variable selection, outlier detection and error correction processes. Fig. 4 show nodal inputs 410 as the average number of cases per year per the selected provider for the basis of scoring the provider mentioned in Granson’s column 11, lines 15-48, column 13, lines 16-49.) and
generate and train a machine learning algorithm based on the plurality of variables associated with physician attribution such that the machine learning algorithm is configured to iteratively determine a set of weighted predictor parameters indicative from the plurality of variables (See surgical quality rating, Surgeon check (column 7, line 35 to column 8, line 26), values objectively represent surgeon qualifications, patient symptom and recovery evaluation criteria, medical facility evaluation criteria including staff, programs and infrastructure, and cost metrics, each of these values having been calculated through the SurgeonCheck ETL 142, also see column 11, lines 15-48. See data set fed in column 17, lines 4-17 where a physician and surgical procedures are scored. Besides predicting decision alternatives in column 25, lines 3-10, see winning model in [column 29, line 66 to column 30, line 14] predictive models utilizing machine learning algorithms for the metric average number of cases per year per provider, a metric from the metric set for evaluating the Cholecystectomy procedure.), wherein all two-way interactions between each variable of the plurality of variables is considered (See Fig. 4, column 10, line 38 to column 11, line 2 relationship of dependent and independent variables.) and wherein the machine learning algorithm is configured to predict a physician attribution score for a physician associated with the patient encounter representative of the physician's degree of involvement in the patient encounter from subsequent EHR data based on the set of weighted predictor parameters (See column 11, lines 26-48, column 17, lines 25-54 scoring a provider is based on the provider having further experience, training and knowledge as a specialist member of the Fellowship of Trained Interventional Radiology and cytopathologist are parameters.).
Although Granson discloses a method of computing physician attribution, using machine learning assigning a physician attribution score representative of a physician’s degree of involvement in the patient encounter for each physician associated with the patient encounter based on a set of predictor parameters and weighting values and improves attribution data analysis mentioned above, Granson does not explicitly teach indicating that each physician’s attribution decision had a higher degree among physicians involvement in the patient encounter. Lieber teaches tracking a degree of physician involvement in a patient encounter (See P0002-P0003, P0016-P0017 tracking Relative Value Units (RVU) as degrees of involvement for physician involvement in a patient encounter mentioned in P0020-P0021, P0025-P0026.) and:
attributed for a historic patient encounter where the one or more physicians had a higher degree of involvement in the historic patient encounter than all other physicians involved in the historic patient encounter (See P0013 and [P0027] Certain embodiments allow the comparison of the information regarding the procedures and the associated RVU or Value Unit to help further analyze the particular physician or health care provider's performance and compensation in relation to other specialties. Certain procedures or services may have different RVUs or Value Units assigned to the same procedure or service, based on the assumption that the procedure or service is more difficult for one particular specialty and not the other. Also, see peer-to-peer comparing and metrics in P0028, and Total Value Unit per physician in Fig. 2, P0059.).
Therefore, it would have been obvious to one of ordinary skill in the art of managing physician value units before the effective filing date of the claimed invention to modify the method, system and software of Granson to include indicating that each physician’s attribution decision had a higher degree among physicians involvement in the patient encounter as taught by Lieber to further refine the value of the service or procedure performed to further facilitate appropriate reimbursement and compensation mentioned in Lieber’s P0002.
Regarding claim 2, Granson discloses wherein the plurality of variables includes a physician type, a physician status, and a length of stay associated with a record of a past encounter with a physician predetermined to be properly attributed (See column 6, lines 21-32 for physicians attributed to performed procedures for lung cancer, Y bypass and a colonoscopy. Also, see education and certification suitability of a physician in column11, lines 15-25, length of stay metric in column 21, line 61 to column 22, line2.).
Regarding claim 5, Granson discloses evaluating accuracy of the machine learning algorithm by feeding the machine learning algorithm with information from a validation set defining additional predefined decisions for physician attribution and associated variables (Besides data feeds in column to third party payers in column 5, lines 25-37, see feed-forward hidden nodes in column 10, lines 18-37. Also, see [column 23, lines 3-17] a competitive scoring routine utilizes this series of validation calculations to statistically analyze the results from each of the machine learning techniques and compare the target values against patterns that were detected in the original data.).
Regarding claim 6, Granson discloses wherein the machine learning algorithm includes a regression algorithm that models relationships between the plurality of variables to extract the set of predictor parameters (See [column 5, line 52 to column 6, line 10] the SurgeonCheck extract-transform-load (ETL) algorithm 142, establishes objective metric data points relevant to the selection of a provider and facility, and prediction of cost for a specified medical procedure. See refining the machine learning algorithm using a measure of error in column 21, lines 12-25 as validation calculations statistically analyzes the results from each of the techniques and compares the generated target values against patterns that were detected in the original data.).
Regarding claim 9, Granson discloses the system of claim 8, wherein the machine learning algorithm is a regression model (See regression algorithm in Fig. 10, 1012 and predictors mentioned in column 24, lines 31-43.).
Regarding claim 10, Granson discloses the system of claim 9, wherein the regression model includes at least one of linear regression, logistic regression, polynomial regression, stepwise regression, ridge regression, LASSO regression, or ElasticNet regression (See linear regression in column 23, lines 60-67.).
Regarding claim 11, Granson discloses the system of claim 9, wherein the processor of the second computing device iteratively modifies a weighting value of each of the weighted predictor parameters until a rate of error for the machine learning algorithm is within a predefined threshold (Taught in column 3, lines 40-47 as exemplary neural network use to minimize statistical errors also mentioned in column 14, lines 35-67. See Fig. 13A-13C, and taught in column 30, lines 15-49, as upper-boundary cutoff and minimum number of sample points of predictors plotted.).
Claims 12 and 15-16 are rejected under 35 U.S.C. 103 as being unpatentable over Granson (US 11,816,539 B1) in view of Szeto (US 2018/0018590 A1) further in view of Lieber (US 2014/0278493 A1).
Claim 12:
Granson discloses a tangible, non-transitory, computer-readable media having instructions encoded thereon, the instructions, when executed by a processor (See Fig. 8, processor 802 and machine-readable medium 822 mentioned in column 18, line 48 to column 19, line 2. See health care provider quality metric values in column 3, lines 32-35. Also see SurgeonCheck in column 6, lines 21-32 for physicians attributed to performed procedures for lung cancer, Y bypass and a colonoscopy.), are operable to:
access EHR data (Besides exemplary dataset of input nodes 410 shown in in Fig. 4, column 11, lines 15-48 representing the clinical history of the physician, see column 5, lines 6-36 record detailing medical staff, medical imaging and historical patient treatment data. In [column 3, lines 18-36] A plurality of input data metrics from the database is input to a corresponding plurality of input nodes operated by one or more processors.); and
generate a training dataset from the EHR data including predefined decisions of physician attribution where one or more physicians were attributed (See machine learning in [column 15, lines 1-40, column 17, lines 4-17] the output data are the results of the neural network predictive model for a metric target value for the average number of cases per year per provider. The neural network process utilizes the optimized variable selection set from the Bayesian network model as the training set input. Besides exemplary dataset of input nodes 410 shown in in Fig. 4, column 11, lines 15-48 representing the clinical history of the physician, see column 5, lines 6-36 record detailing medical staff, medical imaging and historical patient treatment data. In [column 3, lines 18-36] A plurality of input data metrics from the database is input to a corresponding plurality of input nodes operated by one or more processors.),
extrapolate a plurality of variables associated with physician attribution from the training dataset (Besides natural language extraction, extracting a plurality of threshold values or threshold value ranges (column 9, lines 1-29), see [column 20, lines 57-67] these other machine learning algorithms extract imputed data with the automated variable selection, outlier detection and error correction processes. Fig. 4 show nodal inputs 410 as the average number of cases per year per the selected provider for the basis of scoring the provider mentioned in Granson’s column 11, lines 15-48, column 13, lines 16-49.) and
generate and train a machine learning algorithm based on the plurality of variables associated with physician attribution such that the machine learning algorithm is configured to iteratively determine a set of weighted predictor parameters indicative from the plurality of variables (See surgical quality rating, Surgeon check (column 7, line 35 to column 8, line 26), values objectively represent surgeon qualifications, patient symptom and recovery evaluation criteria, medical facility evaluation criteria including staff, programs and infrastructure, and cost metrics, each of these values having been calculated through the SurgeonCheck ETL 142, also see column 11, lines 15-48. See data set fed in column 17, lines 4-17 where a physician and surgical procedures are scored. Besides predicting decision alternatives in column 25, lines 3-10, see winning model in [column 29, line 66 to column 30, line 14] predictive models utilizing machine learning algorithms for the metric average number of cases per year per provider, a metric from the metric set for evaluating the Cholecystectomy procedure.), wherein all two-way interactions between each variable of the plurality of variables is considered (See Fig. 4, column 10, line 38 to column 11, line 2 relationship of dependent and independent variables.) and wherein the machine learning algorithm is configured to predict a physician attribution score for a physician associated with a patient encounter representative of the physician’s degree of involvement in the patient encounter from subsequent EHR data based on the set of weighted predictor parameter (See column 11, lines 26-48, column 17, lines 25-54 scoring a provider is based on the provider having further experience, training and knowledge as a specialist member of the Fellowship of Trained Interventional Radiology and cytopathologist are parameters.).
Although Granson discloses a system for computing physician attribution mentioned above, Granson does not explicitly teach using LASSO regression to select the set of predictor parameters via leave-one- encounter-out cross validation. Szeto teaches:
using a LASSO regression model via leave-one-encounter-out cross validation (See least absolute shrinkage and selection operator (LASSO) and algorithm in P0069. See cross validation in P0098.).
Therefore, it would have been obvious to one of ordinary skill in the art of machine learning before the effective filing date of the claimed invention to modify the method, system and software of Granson to include performing LASSO regression to select the set of predictor parameters via leave-one- encounter-out cross validation as taught by Szeto to benefit from technologies that could extract learned information or “knowledge” while also respecting private or secured information distributed across multiple data stores mentioned in Szeto’s paragraph 5.
Although Granson and Szeto teach computing physician attribution, using machine learning assigning a physician attribution score representative of a physician’s degree of involvement in the patient encounter for each physician associated with the patient encounter based on a set of predictor parameters and weighting values and improves attribution data analysis mentioned above, Granson and Szeto does not explicitly teach indicating that each physician’s attribution decision had a higher degree among physicians involvement in the patient encounter. Lieber teaches tracking a degree of physician involvement in a patient encounter (See P0002-P0003, P0016-P0017 tracking Relative Value Units (RVU) as degrees of involvement for physician involvement in a patient encounter mentioned in P0020-P0021, P0025-P0026.) and:
attributed for a historic patient encounter where the one or more physicians had a higher degree of involvement in the historic patient encounter than all other physicians involved in the historic patient encounter (See P0013 and [P0027] Certain embodiments allow the comparison of the information regarding the procedures and the associated RVU or Value Unit to help further analyze the particular physician or health care provider's performance and compensation in relation to other specialties. Certain procedures or services may have different RVUs or Value Units assigned to the same procedure or service, based on the assumption that the procedure or service is more difficult for one particular specialty and not the other. Also, see peer-to-peer comparing and metrics in P0028, and Total Value Unit per physician in Fig. 2, P0059.).
Therefore, it would have been obvious to one of ordinary skill in the art of managing physician value units before the effective filing date of the claimed invention to modify the method, system, software of Granson and Szeto to include indicating that each physician’s attribution decision had a higher degree among physicians involvement in the patient encounter as taught by Lieber to further refine the value of the service or procedure performed to further facilitate appropriate reimbursement and compensation mentioned in Lieber’s P0002.
Regarding claim 15, Granson discloses a tangible, non-transitory, computer-readable media of claim 12, further comprising additional instructions that when executed by the processor are operable to: apply a regularization method and shrink one more coefficients of a machine learning model to Zero to improve feature selection of a set of parameter-predictors for the machine learning algorithm (See [column 24, lines 16-42] Initially all coefficients are zero, as is the predicted response. The predictor that is most correlated with the current residual is identified, and a step is taken in the direction of this predictor.).
Regarding claim 16, Granson discloses a tangible, non-transitory, computer-readable media of claim 12, further comprising additional instructions that when executed by the processor are operable to: iteratively modify a weighting value of each of the weighted predictor parameters until a rate of error for the machine learning algorithm is within a predefined threshold (See column 22, lines 9-33 where literature-based metric targets serve as predefined threshold. Also, see exemplary threshold value range in column 9, lines 14-29. See column 11, lines 26-48, column 17, lines 25-54 scoring a provider is based on the provider having further experience, training and knowledge as a specialist member of the Fellowship of Trained Interventional Radiology and cytopathologist are parameters.).
Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over Granson (US 11,816,539 B1) in view of Lieber (US 2014/0278493 A1) and Chen (US 11,398,299 B2).
Regarding claim 3, although Granson and Lieber teach the set of predictor parameters as mentioned above, Granson and Lieber do not explicitly teach additional predictor parameters. Chen teaches wherein the set of predictor parameters includes daily progress notes per day, long notes per day, orders per day, a length of stay, and an attending parameter (See medical notes in Abstract, attending to patients in column 1, lines 35-45 and 51-55, column 2, see ordering lab tests, administering glucose daily in lines column 19, lines 51-54. See readmission and length of stay in column 18, lines 24-30.).
Therefore, it would have been obvious to one of ordinary skill in the art of summarizing medical events before the effective filing date of the claimed invention to modify the method, system and software of Granson and Lieber to include the set of predictor parameters includes daily progress notes per day, long notes per day, orders per day, a length of stay, and an attending parameter as taught by Chen to allocate their attention efficiently among the overabundance of information from diverse sources, as well as to provide predictions of future clinical events mentioned in Chen’s column 11, lines 45-55.
Claims 4 are rejected under 35 U.S.C. 103 as being unpatentable over Granson (US 11,816,539 B1) in view of Lieber (US 2014/0278493 A1) and Szeto (US 2018/0018590 A1).
Regarding claim 4, although Granson and Lieber teach the set of predictor parameters as mentioned above, , Granson and Lieber do not explicitly teach additional predictor parameters. Szeto teaches performing LASSO regression to extract the set of predictor parameters via leave-one- encounter-out cross validation (See least absolute shrinkage and selection operator (LASSO) and algorithm in P0069. See cross validation in P0098.).
Therefore, it would have been obvious to one of ordinary skill in the art of machine learning before the effective filing date of the claimed invention to modify the method, system and software of , Granson and Lieber to include performing LASSO regression to select the set of predictor parameters via leave-one- encounter-out cross validation as taught by Szeto to benefit from technologies that could extract learned information or “knowledge” while also respecting private or secured information distributed across multiple data stores mentioned in Szeto’s paragraph 5.
Claims 13 is rejected under 35 U.S.C. 103 as being unpatentable Granson (US 11,816,539 B1) in view of Szeto (US 2018/0018590 A1) further in view of Lieber (US 2014/0278493 A1) and Park (US 2013/0151516 A1).
Regarding claim 13, Granson, Szeto and Lieber do not explicitly teach casual effect relationships between variables. Park teaches wherein the machine learning algorithm executed by the processor finds casual effect relationships between variables of the plurality of variables (See [P0055] The regression analysis is a statistical method for predicting casual relationship between a plurality of variables.).
Therefore, it would have been obvious to one of ordinary skill in the art of clinical data analysis before the effective filing date of the claimed invention to modify the method, system and software of Granson, Szeto and Lieber to include casual effect relationships between variables as taught by Park to give attention to dangerous, disease progression as mentioned in Park’s P0060-P0061.
Response to Arguments
Applicant argues that the Office generalizes steps within "training" the machine learning model or associated with the machine learning algorithm. see pgs. 8-9 of Remarks – Examiner disagrees.
No machine learning process is claimed. Instead, unknown physician attributes take the form of variables extracted datasets from electronic health records, with no criteria for determining “a higher degree of involvement”, further predicting and assigning weights, where the unknown physician attributes remain the same throughout data processing steps of claims 1, 8 and 12. No data is trained as a resulting of a machine learning process.
Furthermore, the functional additional element steps of training a machine learning algorithm through extraction of predictor parameters via leave-one-encounter-out where all two- way interactions are considered, assigning weighting values to the predictor parameters, and iterative modification of the weighting values such that the trained algorithm can predict a physician attribution score amount to insignificant extra-solution activity and are well-understood, routine, and conventional in the art, as evidenced by at least column 14, lines 6-17, column 7, line 35 to column 8, line 26, column 9, lines 1-29, column 11, lines 15-48, column 17, lines 4-17, column 20, lines 57-67, column 25, lines 3-10, and column 29, line 66 to column 30, line 14 of Granson (US 11,816,539 B1). Abstract, Fig. 2, 0006-P0007, 0048-P0049 of Ridgeway (WO 2018/075945 A1) et al. The functional additional elements, as amended, the steps of assigning a weighting value to learned predictor parameters and modifying the weighting value of the gathered predictor parameters using a machine learning algorithm within a predefined threshold are well-understood, routine, and conventional in the art, as evidenced by at least column 7, line 35 to column 8, line 26, column 9, lines 1-29, column 11, lines 15-48, column 17, lines 4-17, column 20, lines 57-67, column 25, lines 3-10, and column 29, line 66 to column 30, line 14 of Granson (US 11,816,539 B1).
Regarding the prior art rejections, Applicant’s arguments have been fully considered, but are now moot in view of the new grounds of rejection. The Examiner has entered a new rejection under 35 USC § 103(a) and applied new art and art already of record.
Applicant’s arguments, see page 9, filed 12/22/2025, with respect to 112 rejection have been fully considered and are persuasive. Applicant’s amended claim language is persuasive and addresses the 112 rejection of claims 1-6, 8-13 and 15-16, which has been withdrawn.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to TERESA S WILLIAMS whose telephone number is (571)270-5509. The examiner can normally be reached Mon-Fri, 8:30 am -6:30 pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Mamon Obeid can be reached at (571) 270-1813. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/T.S.W./Examiner, Art Unit 3687 04/04/2026
/ALAAELDIN M. ELSHAER/Primary Examiner, Art Unit 3687